
Memory-injection
Research on persistent prompt injections via memory features in Claude, Gemini, and ChatGPT, with reproduction steps, datasets, and proofs for…

Research on persistent prompt injections via memory features in Claude, Gemini, and ChatGPT, with reproduction steps, datasets, and proofs for…

Research implementation of the Black-Hole Attack that poisons vector database indexes via centrality-driven hubness, with FAISS pipelines and…

Zero-retraining sanitization that reorders permutation-equivalent neural network units to destroy malware embedded via weight-steganography attacks…

Research pipeline that constructs verified successful experiences and organizes them into formation records to progressively poison model skills via…

Defender framework for LLM agent security that compiles task contracts, validates capability manifests, and checks effects via PLANT/WRAP proof…

Code, response logs and labels for the paper Jailbreaking Open-Weight LLMs via Random Embedding Perturbations

Embeds and detects keyed, style-based watermarks in LLM-generated Python, Java, and C++ code via CST rewriting, preserving functional correctness and…

Standard-library Python security triage engine that scans web apps, APIs, LLMs, and mobile packages via passive header inspection, active canary…

Evolutionary LLM jailbreak and guardrail framework that grows a reusable strategy pool via genetic mutation, Markov selection, and online adversarial…

AI-assisted research pipeline that extracts HTTP desync techniques, generates malformed request test-cases, validates them via Burp, and confirms…

Black-box input-stage purification defense that neutralizes backdoor attacks on object detectors via corruption, diffusion reconstruction, and DBSCAN…

Defensive framework that maintains a safety-focused shadow memory to detect and block prompt-injection and long-horizon threats against LLM agents…

Proof-of-concept that poisons MLflow registered models via the REST API, embedding a malicious pickle to trigger RCE when the model is loaded.

Security research lab — Unauthenticated RCE via insecure deserialization in ComfyUI v0.23.0 (CVSS 9.8). Isolated Docker environment, technical…

Natural-language Android automation agent that drives real devices via ADB, captures Logcat and screenshots, and exposes an MCP server for AI IDEs…

Safety cannot be a prompt instruction. TBP provides an external execution-layer boundary for autonomous agents, enforcing hard F/I/W invariants via…

Syntactic Ghost: An Imperceptible General-purpose Backdoor Attacks on Pre-trained Language Models

AI red-team platform. Autonomous LLM agents run a penetration test end to end inside a Kali container and write the report. LangGraph plan/act…